Carthagène A brief introduction to combinatorial optimization: The Traveling Salesman Problem Simon de Givry.

Carthagène

A brief introduction to combinatorial optimization:

The Traveling Salesman Problem

Simon de GivrySimon de Givry

Find a tour with minimum distance, visiting every city only once

Distance matrix (miles)Distances Camara Caniço Funchal ...

Camara 0 15 7 ...

Caniço 15 0 8 ...

Funchal 7 8 0 ...

... ... ... ... ...

Find an order of all the markers with maximum likelihood

2-point distance matrix (Haldane)

Distances M1 M2 M3 ...

M1 0 14 7 ...

M2 14 0 8 ...

M3 7 8 0 ...

... ... ... ... ...

Link

M1 M2 M3 M4 M5M6 M7

Mdummy

0 0…0…

i,j,k distance(i,j) ? distance(i,k) + distance(k,j)=,

Multi-point likelihood (with unknowns) the distance between two markers depends on the order

Traveling Salesman Problem

Complete graph Positive weight on every edge

Symmetric case: dist(i,j) = dist(j,i) Triangular inequality: dist(i,j) dist(i,k) +

dist(k,j)

Euclidean distance Find the shortest Hamiltonian cycle

78

15

Total distance = xxx miles

Traveling Salesman Problem Theoretical interest

NP-complete problem 1993-2001: +150 articles about TSP

in INFORMS & Decision Sciences databases

Practical interest Vehicle Routing Problem Genetic/Radiated Hybrid Mapping

Problem NCBI/Concorde, Carthagène, ...

Variants Euclidean Traveling Salesman Selection Problem Asymmetric Traveling Salesman Problem Symmetric Wandering Salesman Problem Selective Traveling Salesman Problem TSP with distances 1 and 2, TSP(1,2) K-template Traveling Salesman Problem Circulant Traveling Salesman Problem On-line Traveling Salesman Problem Time-dependent TSP The Angular-Metric Traveling Salesman Problem Maximum Latency TSP Minimum Latency Problem Max TSP Traveling Preacher Problem Bipartite TSP Remote TSP Precedence-Constrained TSP Exact TSP The Tour Cover problem ...

Plan Introduction to TSP Building a new tour Improving an existing tour Finding the best tour

Building a new tour

Nearest Neighbor heuristic

Greedy (or multi-fragments) heuristic

Savings heuristic (Clarke-Wright 1964)

Heuristics Mean distance to the optimum

Savings: 11%

Greedy: 12%

Nearest Neighbor: 26%

Improving an existing tour

Which local modification can improve this tour?

Remove two edges and rebuild another tour

Invert a given sequence of markers

2-change

Remove three edges and rebuild another tour (7 possibilities)

Swap the order of two sequences of markers

« greedy » local search 2-opt

Note: a finite sequence of « 2-change » can reach any tour, including the optimum tour

Strategy: Select the best 2-change among N*(N-1)/2

neighbors (2-move neighborhood) Repeat this process until a fix point is

reached (i.e. no tour improvement was made)

2-opt

Greedy local search Mean distance to the optimum

2-opt : 9% 3-opt : 4% LK (limited k-opt) : 1%

Complexity 2-opt : ~N3

3-opt : ~N4

LK (limited k-opt) : <N4 ?

Complexity n = number of vertices

Algorithm Complexity A-TSP (n-1)! S-TSP (n-1)! / 2 2-change 1 3-change 7 k-change (k-1)! . 2k-1 k-move (k-1)! . 2k-1 . n! / (k! . (n-k)!) ~ O( nk ) k << n

In practice: o( n ) 2-opt et 3-opt ~ O( nk+1 )

In practice: o( n1.2 ) time(3-opt) ~ 3 x time(2-opt)

For each edge (uv), maintain the list of vertices wsuch that dist(w,v) < dist(u,v)

u

v

2-opt implementation trick:

Lin & Kernighan (1973) k-change : e1->f1,e2->f2,...

=> Sumki=1( dist(ei) - dist(fi) ) > 0

There is an order of i such that all the partial sums are positives:

Sl = Sumli=1( dist(ei) - dist(fi) ) > 0

=> Build a valid increasing alternate cycle:xx ’->yx ’, yy’ -> zy’, zz’ -> wz’, etc.dist(f1)<dist(e1),dist(f1)+dist(f2)<dist(e1)+dist(e2),..+ Backtrack on y and z choices + Restart

x

x’

y

y’

z

z’

w

(in maximization)

e2

e3

e4

e1

f1

f3

f2

f4

w’

{x,y,z,w,..} ^ {x’,y’,z’,w’,..} = 0y is among the 5 best neighbors of x’, the same for z’ and w

Is this 2-opt tour optimum?

2-opt + vertex reinsertion

local versus global optimum

Local search &« meta-heuristics » Tabu Search

Select the best neighbor even if it decreases the quality of the current tour

Forbid previous local moves during a certain period of time

List of tabu moves Restart with new tours

When the search goes to a tour already seen

Build new tours in a random way

Tabu Search

• Stochastic size of the tabu list• False restarts

Experiments with CarthaGèneN=50 K=100 Err=30% Abs=30%

Legend: partial 2-opt = early stop , guided 2-opt 25% = early stop & sort with X = 25%

Experiments - next

Other meta-heuristics Simulated Annealing

Local moves are randomly chosen Neighbor acceptance depends on its quality

Acceptance process is more and more greedy Genetic Algorithms

Population of solutions (tours) Mutation, crossover,…

Variable Neighborhood Search …

Simulated AnnealingMove from A to A’ acceptedif cost(A’) ≤ cost(A)or with probability P(A,A’) = e –(cost(A’) – cost(A))/T

Variable Neighborhood Search

• Perform a move only if it improves the previous solution• Start with V:=1. If no solution is found then V++ else V:=1

Local Search

Demonstration

Finding the best tour

Search tree

M2M1 M3

M2 M3 M1 M3 M1 M2

M3 M2 M3 M1 M2 M1

M1,M2,M3

depth 1:

depth 2:

depth 3:

leaves

node

branch

root

= choice point

= alternative

= solutions

Tree search Complexity : n!/2 different orders

Avoid symmetric orders (first half of the tree)

Can use heuristics in choice points to order possible alternatives

Branch and bound algorithm Cut all the branches which cannot lead to a

better solution

Possible to combine local search and tree search

Branch and boundMinimum weight spanning tree

Prim algorithm (1957)

Held & Karp algorithm (better spanning trees) (1971) linear programming relaxation of TSP, LB(I)/OPT(I) 2/3

Christofides heuristic (1976)

=> A(I) / OPT(I) 3/2 (with triangular inequalities)

Complexity

Complexity

Standard computer

Computer 100 times faster

Computer 1000 times faster

N N1 100*N1 1000*N1

N2 N2 10*N2 31,6*N2

N3 N3 4,64*N3 10*N3

2N N4 N4+6,64 N4+9,97

3N N5 N5+4,19 N5+6,29

Complete methods 1954 : 49 cities 1971 : 64 cities 1975 : 100 cities 1977 : 120 cities 1980 : 318 cities 1987 : 2,392 cities 1994 : 7.397 cities 1998 : 13.509 cities 2001 : 15.112 cities (585936700 sec. 19 years of CPU!)